My hobbies are to code a toy program, to play a guitar and to enjoy Starcraft. I love traveling with my wife to awesome places.
I am working as a research engineer in Samsung Research Artificial Intelligence Center. I received master's dgree under Prof. Jaegul Choo in the Department of Big Data Convergence at Korea University. I earned bachelor's degree under Prof. Jeehyong Lee in the Department of Computer Science and Engineering at Sungkyunkwan University.
Most interested in deep-learning, especially for natural language processing. In detail, my research mainly focuses on building algorithms to teach machines how to read and comprehend written texts.
Keywords: AI, Machine Learning, Deep Learning, NLP, MRC, Text SummarizationDownload CV
- Dec 2019 — One paper has been accepted at ICLR 2020.
- Oct 2019 — I received Bronze(4th) prizes at Samsung Best Paper Award 2019 (AI division).
- SEP 2019 — Our EPS model is at the top of the HotpotQA-fullwiki (EM score) leaderboard.
- JUN 2019 — Our BERT-CLKT (with ASGen) is at the top of the KorQuAD-leaderboard.
- May 2019 — Our EPS model is at the top of the HotpotQA-distractor leaderboard.
- Oct 2018 — I received Gold(1st), Silver(2nd), and Bronze(4th) prizes at Samsung Best Paper Award 2018 (AI division).
- Aug 2018 — Two papers have been accepted at EMNLP 2018.
- Jul 2018 — My works and I appeared in two articles "With Samsung’s Unique Strengths, We Are Developing a User-Oriented AI Algorithm", "Samsung Electronics Wins at Two Top Global AI Machine Reading Comprehension Challenges".
- Jul 2018 — Our ConZNet is at the top of the MS MARCO leaderboard.
- Jun 2018 — Our MemoReader is at the top of the TriviaQA-Web (F1 score) leaderboard.
- May 2018 — One paper has been accepted at ACL 2018 workshop on MRQA.
- Feb 2018 — Our MAMCN+ is ranked 5th (single model) in the SQuAD v1.1 leaderboard.
PublicationsSee more detail in Google Scholar
NeurQuRI: Neural Question Requirement Inspector for Answerability Prediction in Machine Reading Comprehension
Seohyun Back, Sai Chetan Chinthakindi, Akhil Kedia, Haejun Lee and Jaegul Choo
Conference on Learning Representations (ICLR), 2020, Addis Ababa, Ethiopia
MemoReader: Large-Scale Reading Comprehension through Neural Memory Controller
Seohyun Back, Seunghak Yu, Sathish Indurthi, Jihie Kim and Jaegul Choo
Conference on Empirical Methods in Natural Language Processing (EMNLP), 2018, Brussels, Belgium, Long paper
Cut to the Chase: A Context Zoom-in Network for Reading Comprehension
Sathish Indurthi, Seunghak Yu,Seohyun Back and Heriberto Cuayahuitl
Conference on Empirical Methods in Natural Language Processing (EMNLP), 2018, Brussels, Belgium, Short paper
A Multi-Stage Memory Augmented Neural Network for Machine Reading Comprehension
Seunghak Yu, Sathish Indurthi,Seohyun Back and Haejun Lee
Proceedings of the Workshop on MRQA, Melbourne, Australia, 2018. Association for Computational Linguistics (ACL)
Text Summarization using Self-Attention and Self-Learning
Jisang Yu, Seohyun Back and Jaegul Choo
Conference on Korea Software Congress (KSC2018), 2018, Pyeongchang, South Korea
WAS Framework Designed for Distributed Computing Environments
Ungueop Choi, Eunji Kwon, Jeongsub Lee and Seohyun Back
Proceedings of The Korea Contents Association (KoCon) Spring Conference, 2013
Document Summarization Using Mutual Recommendation with LSA and Sense Analysis
Seohyun Back, Dongwook Lee, Minji Park, Jinhee Park, Hyewuk Jung and Jeehyong Lee
Proceedings of Korean Institute of Intelligent Systems (KIIS) Spring Conference 2012 (Best Paper Award)
Moving Pattern based Bot Detection Model
Seohyun Back, Jaekwang Kim and Jeehyong Lee
Proceedings of Korean Institute of Intelligent Systems (KIIS) Fall Conference 2011
ProjectsSee more detail in GitHub
- Bianca (on going) — Currently I am making a framework for personal assistant agent. This project (will) provide serveral natural language understanding models and conversation models.
- Bixby — Voice assistant. Developed a rejection module to exclude non-processable commands.
- Commercial chatbot — Developed chitchat model, crawler, dialog manager and customized intent classifier.
- Blueeye — Indoor pointing system. Customized GMM algorithm to predict indoor pointing using mobile device based oncollected RF signal.
- Pytrain — Machine Learning library for python. While doing research on indoor pointing system, I developed own machine learning library from the scratch.
- Wikilite — Personal markdown based minimal blog & wiki engine.
- Machine translation — Integrated ASR, MT, TTS modules to make end-to-end system and built PoC.
- Dialog summarization — Integrated ASR, dialog manager modules to make phone call summarization.
- Friends take this — Web market framework. Developed p2p market solution that consist of web server and Android application.
- Athene square — Distributed streaming system. Developed web browser p2p streaming system using web-rtc protocol.
- Smart Highlighter — Document summarization solution. Developed a tool for document analysis and extractive summarization.